Fda postmarket surveillance-report-september-remediated

Strengthening
Our natiOnal SyStem
fOr medical device
POStmarket
Surveillance

center fOr deviceS and radiOlOgical health
u.S. fOOd and drug adminiStratiOn

SEPTEMBER 2012

table Of cOntentS

a InTRoducTIon 03

B cuRREnT Fda MEdIcal dEvIcE PoSTMaRkET 05
SuRvEIllancE SySTEM
1 Medical device RepoRting (MdR) 05
2 Medical pRoduct Safety netwoRk (MedSun) 05
3 poSt-appRoval StudieS 06
4 poStMaRket SuRveillance StudieS (“522 StudieS”) 06
5 fda diScRetionaRy StudieS 06
6 otheR toolS 06

c STREngThEnIng ouR naTIonal SySTEM FoR 08

MEdIcal dEvIcE PoSTMaRkET SuRvEIllancE

1 eStabliSh a unique device identification SySteM and pRoMote itS 10

incoRpoRation into electRonic health infoRMation

2 pRoMote the developMent of national and inteRnational device 10

RegiStRieS foR Selected pRoductS

3 ModeRnize adveRSe event RepoRting and analySiS 13

3.1 Development of AutomAteD ADverse event reporting systems 13

3.2 increAse the number of mDrs receiveD electronicAlly 14

3.3 Develop A mobile ApplicAtion for ADverse event reporting 14

3.4 moDernize the meDicAl Device ADverse event DAtAbAse 14

3.5 rApiDly iDentify sAfety signAls 15

4 develop and uSe new MethodS foR evidence geneRation, 16

SyntheSiS and appRaiSal

4.1 QuAntitAtive Decision AnAlysis to evAluAte benefits AnD risks 16

4.2 eviDence Assessment by combining DAtA from Diverse DAtA sources 17

4.3 AutomAteD signAl Detection 18

4.4 refinement of processes for signAl Detection AnD mAnAgement 18

d concluSIon 21

E aPPEndIx 22

current selecteD meDicAl Device postmArket Authorities

Strengthening our national system for medical device postmarket surveillance 02

a. intrOductiOn

Several high-profile medical device perfor we strive to permit marketing of only those
mance concerns have led some to question devices with a favorable benefit-risk profile,
whether the current United States post- even a thorough premarket product evalua
market surveillance system is optimally tion can leave some unanswered questions
structured to meet the challenges of rapidly about a medical device’s performance and
evolving medical devices and the changing associated clinical benefits and risks.
nature of health care delivery and informa
tion technology. In their report entitled, We believe that strengthening our National
“Medical Devices and the Public’s Health: Medical Device Surveillance System will
The FDA 510(k) Clearance Process at 35 complement the improvements we have
Years,” published in July 2011, the Institute made to our premarket program. A medi
of Medicine (IOM) recommended that the cal device postmarket surveillance system
Food and Drug Administration (FDA) de should quickly identify poorly performing
velop and implement a comprehensive med devices, accurately characterize and dissem
ical device postmarket surveillance strategy inate information about real-world device
to collect, analyze, and act on medical de performance, including the clinical benefits
vice postmarket performance information. and risks of marketed devices, and efficient
ly generate data to support premarket clear
Medical device product evaluation presents ance or approval of new devices and new
unique challenges compared to drugs and uses of currently marketed devices.
biologics, related to the greater diversity and
complexity of medical devices, and the rapid This document, in addition to providing
technological advances and iterative nature an overview of FDA’s medical device post-
of medical device product development. market authorities and the current United
FDA's Center for Devices and Radiological States medical device postmarket surveil
Health (CDRH) has undertaken a number lance system, proposes four specific actions,
of actions since early 2011 to strengthen our using existing resources and under current
premarket review program and facilitate the authorities, to strengthen the medical de
timely delivery of innovative, safe and effec vice postmarket surveillance system in the
tive products to American patients.1 While United States.

1 A description of the actions cDrh has taken can be found at: http://www.fda.gov/AboutfDA/centersoffices/
officeofmedicalproductsandtobacco/cDrh/cDrhreports/ucm239448.htm.


These actions are: We are suggesting smart, tailored modi
fications to the existing postmarket
1. Establish a Unique Device Identification surveillance program. Furthermore, we
System and Promote Its Incorporation recognize that our postmarket vision
into Electronic Health Information; cannot be implemented or achieved by
the FDA alone. It requires the input
2. Promote the Development of National and active participation of many other
and International Device Registries for key domestic and foreign stakehold
Selected Products; ers including the medical device in
dustry, health care providers, patients,
3. Modernize Adverse Event Reporting academia, third-party payers, hospitals
and Analysis; and, and other health care facilities, health
care data holders, and other government
4. Develop and Use New Methods for agencies. In addition, CDRH is hosting
Evidence Generation, Synthesis and four public meetings in September 2012
Appraisal. to garner stakeholder feedback.


current fda medical device POStmarket

b. Surveillance SyStem

Postmarket surveillance is the systematic and scientific and medical literature. The
collection, analysis, interpretation, and current United States medical device post-
dissemination of health-related data to im market surveillance system depends primar
prove public health and reduce morbidity ily upon:
and mortality. Medical device postmarket
surveillance presents unique challenges 1. Medical Device Reporting (MDR) —
compared to drugs and biologics due to the Each year, the FDA receives several
great diversity and complexity of medi hundred thousand medical device re
cal devices, the iterative nature of medical ports of confirmed or possible device-
device product development, the learning associated serious injuries, deaths,
curve associated with technology adoption, and malfunctions. While MDRs are
and the relatively short product life cycle. a valuable source of information, this
Proper medical device operation depends passive surveillance system has notable
on optimal device design, the use environ limitations, including the potential
ment, user training, and adherence to direc submission of incomplete or inaccurate
tions for use and maintenance. In some cas data, under-reporting of events, lack of
es, these features limit the utility of relying denominator (exposure) data, and the
on systems designed for the identification of lack of report timeliness.
drug-related adverse events.
2. Medical Product Safety Network
Under its existing authorities, the FDA 2
(MedSun) — MedSun is an enhanced
uses a multifaceted postmarket surveillance surveillance network comprised of ap
approach that relies on various methods proximately 280 hospitals nationwide
and techniques tailored to the specific de that work interactively with the FDA to
vice and public health need. This leverages a better understand and report on device
variety of data sources to monitor the safety use and adverse outcomes in the real-
and effectiveness of marketed medical de world clinical environment. The overall
vices, including repositories of spontane quality of the approximately 5,000
ous reports, device registries, administra reports received annually via MedSun is
tive and claims data, data from integrated significantly higher than those received
health systems, electronic health records, via MDR. Specialty networks within

2 see Appendix for a description of key, current fDA medical device postmarket authorities.


MedSun focus on device-specific areas 5. FDA Discretionary Studies — In ad
such as cardiovascular devices (Heart- dition to medical device adverse event
Net) and pediatric intensive care unit reports, post-approval and postmarket
devices (KidNet). In addition, the net surveillance studies, the FDA also
work can be used for targeted surveys conducts its own research to moni
and focused clinical research. tor device performance, investigate
adverse event signals and characterize
3. Post-Approval Studies — The FDA may device-associated benefits and risks to
order a post-approval study as a condition patient sub-populations. A variety of
of approval for a device approved under a privacy-protected data sources are used
premarket approval (PMA) order. Typi including national registries, Medicare
cally, post-approval studies are used to as and Medicaid administrative and
sess device safety, effectiveness, and/or re claims data, data from integrated health
liability including longer-term, real-world systems, electronic health records, and
device performance. Status updates for published scientific literature.
the more than 160 ongoing post-approval
studies may be found on our website at 6. Other Tools — The FDA has other tools
http://www.accessdata.fda.gov/scripts/ it may use in the postmarket setting
cdrh/cfdocs/cfPMA/pma_pas.cfm. to track devices, restrict or ban device
use, and remove unsafe, adulterated, or
4. Postmarket Surveillance Studies — misbranded products from the market
The FDA may order a manufacturer of (see Appendix).
certain Class II or Class III devices to
conduct postmarket surveillance stud
ies (often referred to as “522 studies”).3
Study approaches vary widely and may
include non-clinical device testing,
analysis of existing clinical databases, ob
servational studies, and, rarely, random
ized controlled trials. Status updates for
ongoing postmarket surveillance studies
covering approximately a dozen device
types may be found on our website at
http://www.accessdata.fda.gov/scripts/
cdrh/cfdocs/cfPMA/pss.cfm.

3 section 522 of the food, Drug & cosmetic Act (fD&c Act).


rOle Of Sentinel initiative in medical approach is enabling the maintenance of patient
device POStmarket Surveillance privacy by keeping directly identifiable patient in
formation behind local firewalls in its existing pro
The Food and Drug Administration Amend tected environment. Additionally, each health care
ments Act of 2007 (FDAAA)4 required the FDA data system retains physical and operational con
to collaborate with public, academic and private trol over its own data, and provides important in
entities to develop methods for obtaining access to put into valid use and interpretation of the data.
disparate health care data sources and to analyze
health care safety data. In May 2008, the Secretary The current Sentinel data model focuses on que
of Health and Human Services (HHS) and FDA’s rying administrative and claims data maintained
Commissioner announced the Sentinel Initiative, by partner organizations, who share aggregated
a long-term effort to create a national electron results with the FDA. Direct identifiers, such as
ic system for monitoring FDA-regulated medical names, are removed from records before infor
product safety. The Sentinel System will ultimate mation is shared with the FDA. De-identified
ly expand FDA’s existing postmarket safety sur information includes outpatient pharmacy data,
veillance systems by enabling the FDA to conduct diagnoses and procedures, mortality, and select
active surveillance and related observational stud laboratory results. The FDA does not receive or
ies on the safety and performance of its regulated hold personally identifiable information, but
medical products once they reach the market. can query privacy-protected data and receive ag
gregated data from local environments that to
The Sentinel Initiative requires the FDA to collab gether total approximately 126 million patients.
orate with public, academic and private entities to Unfortunately, most of these records lack manu
develop methods for obtaining access to disparate facturer or brand-specific device identifiers and
health data sources and to validate means of link therefore cannot be leveraged to perform mean
ing and analyzing health care safety data from those ingful medical device postmarket surveillance.
sources. With this effort, the FDA seeks to create
a privacy-protected, scalable, efficient, and sustain As a result, complementary efforts are required
able system—the Sentinel System—that leverages to develop a comprehensive postmarket surveil
existing electronic health care data from multiple lance system for medical devices. The Food and
sources to actively monitor the safety of regulat Drug Administration Safety and Innovation Act
ed medical products. A key finding from the early of 2012 explicitly requires expansion of FDA’s
work of the Sentinel Initiative is that a distributed Sentinel System to include medical devices.
data system is the preferred approach for perform
ing active surveillance. A distributed system allows The FDA envisions using similar distributed data
data to be maintained in local environments by cur sources, which would include EHRs and regis
rent owners, as opposed to using a centralized ap tries, for the medical device Sentinel System to
proach, which would consolidate the data into one help maintain patient privacy, administrative and
physical location. A key benefit of the distributed claims databases, and other external data sources.

4 public law 110-85.


Strengthening Our natiOnal medical device POStmarket

c. Surveillance SyStem

FDA’s vision for medical device postmar • Reduce burdens and costs of medical
ket surveillance is the creation of a national device postmarket surveillance; and,
system that conducts active surveillance
in near real-time using routinely collected • Facilitate the clearance and approval of
electronic health information containing new devices, or new uses for existing
unique device identifiers, quickly identifies devices.
poorly performing devices, accurately char
acterizes the real-world clinical benefits and The FDA believes that four key steps are
risks of marketed devices, and facilitates the needed to strengthen medical device post-
development of new devices and new uses market surveillance in the United States (see
of existing devices through evidence gen Fig., p. 9). These steps are:
eration, synthesis and appraisal. The system
leverages privacy-protected distributed data 1. Establish a Unique Device Identification
systems and common data standards, and (UDI) System and Promote Its Incorpora
would augment, not replace, other mecha tion into Electronic Health Information;
nisms of surveillance such as FDA’s MDR
and MedSun. 2. Promote the Development of National
and International Device Registries for
Specifically, medical device postmarket sur Selected Products;
veillance should:
3. Modernize Adverse Event Reporting
• Provide timely, accurate, systematic, and and Analysis; and,
prioritized assessments of the benefits
and risks of medical devices throughout 4. Develop and Use New Methods for
their marketed life using high qual Evidence Generation, Synthesis and
ity, standardized, structured, electronic Appraisal.
health-related data;

• Identify potential safety signals in near
real-time from a variety of privacy-
protected data sources;


Figure. Strengthening Our National System for Medical Device Postmarket Surveillance.

The current medical device postmarket surveillance system in the United States depends primarily upon reporting
of possible device-associated serious injuries, deaths and malfunctions (Medical Device Reporting – MDR), an
enhanced surveillance network of approximately 280 hospitals (Medical Product Safety Network – MedSun), studies
ordered by the FDA for selected devices (Post-Approval Studies and Postmarket Surveillance Studies), FDA research
using other data sources (FDA Discretionary Studies), and other tools such as device tracking. The FDA suggests
four specific actions that could be taken using existing resources and under current authorities to strengthen the
medical device postmarket surveillance system in the United States. These actions are: 1) Establish a Unique Device
Identification System (UDI) and Promote Its Incorporation into Electronic Health Information (EHI); 2) Promote
the Development of National and International Device Registries for Selected Products; 3) Modernize Adverse Event
Reporting and Analysis; and, 4) Develop and Use New Methods for Evidence Generation, Synthesis and Appraisal.


1. eStabliSh a uniQue device the device’s characteristics (e.g., whether it
identificatiOn SyStem and PrOmOte contains latex or is magnetic resonance im
itS incOrPOratiOn intO electrOnic aging compatible) and improve the clini
health infOrmatiOn cians ability to trace the device through the
supply chain to the point of patient use.
FDAAA directed the FDA to promulgate
regulations establishing a UDI system for The incorporation of medical device iden
all medical devices. In July 2012, the FDA tifiers into EHRs is another key step that
issued a proposed rule for a UDI system.5 would improve patient safety, make the
A UDI may contain two types of informa conduct of postmarket surveillance more
tion: a unique numeric or alphanumeric efficient, and make queries of and de-iden
code, specific to a device model, and an tified responses from electronic health in
identifier that includes the production in formation more readily usable to support
formation for that specific device, such as device approval or clearance. Likewise, in
the manufacturing lot or batch number, the corporation of UDIs into claims data would
serial number, manufacturing date, and ex increase the utility of these data sources for
piration date. medical device postmarket surveillance, and
pilot studies suggest it is both technically
UDIs will enhance postmarket surveillance feasible and cost-effective.
activities by providing a standard and un
ambiguous way to document device use in In short, a UDI system can improve the de
EHRs, clinical information systems, and tection of medical device adverse events and
claims data sources. As a result, this infor product problems, enhance assessments of
mation would potentially become avail device benefit-risk profiles, streamline and
able for use in assessing the benefits and help secure the domestic and global supply
risks of medical devices. UDIs will also al chain, facilitate more efficient and effective
low the FDA, the health care community recalls, and facilitate the premarket evalu
and industry to more accurately report and ation of new devices and new uses of cur
analyze device-related adverse events by rently marketed devices.
ensuring that critical device information is
included in the reports. These device identi 2. PrOmOte the develOPment Of
fiers may also help reduce medical errors by natiOnal and internatiOnal device
enabling health care professionals and oth regiStrieS fOr Selected PrOductS
ers to rapidly and precisely identify a device,
obtain important information concerning A registry is a system that collects and

5 https://www.federalregister.gov/articles/2012/07/10/2012-16621/unique-device-identification-system


maintains structured records on a specific dis Data Registry or the Society of Thoracic

ease, condition, procedure, or medical prod Surgeon’s Adult Cardiac Surgery Database.

uct for a specified time period and population.
Product registries include patients who have Importantly, the FDA is not seeking to de
been exposed to a specific medical device, velop a centralized repository of registry
biologic or drug product. Health services reg data. Rather each registry should retain
istries consist of patients who have had a com physical and operational control over its
mon procedure, clinical encounter or hospi own data, and provide important input into
talization. Disease or condition registries are valid use and interpretation of its data. This
defined by patients having the same diagnosis, also assures maintenance of patient privacy
such as diabetes mellitus or heart failure. by keeping directly identifiable patient in
formation behind local firewalls in its exist
Procedure and device registries are often ing protected environment. In addition, the
created and maintained by private organi FDA envisions continuing to help facilitate
zations, such as the American College of the creation of registries. It is not seeking to
Cardiology and its National Cardiovascular regulate standards, such as business models

bOx c-1.
uSing the natiOnal SyStem fOr medical device POStmarket Surveillance
tO identify new uSeS Of exiSting deviceS and facilitate market acceSS
fOr innOvative PrOductS

In addition to quickly identifying poorly performing devices and accurately characterizing
the real-world clinical benefits and risks of marketed devices using routinely collected
electronic health information, the National Medical Device Postmarket Surveillance
System should facilitate the development of new technologies, new devices and new uses of
currently marketed devices through evidence generation and analysis. The proposed system
could do so by producing data to:

• Serve as the comparison group or “control arm” in scientific studies evaluating device
performance;
• Identify new patient populations that benefit from device therapy;
• Be leveraged for expansion of labeled device indications to new groups; and,
• Demonstrate the relative safety of a device type to support downclassification and a
reduction in the premarket evidentiary needs.


or taxonomy, for registries. The FDA cur registries that address every medical device
rently partners with and uses registries to as problem or issue. In addition, registry devel
sess the real-world performance of medical opment and maintenance can be associated
products and procedures, to determine the with significant costs and effort. For this rea
clinical effectiveness and safety of a medi son, the creation of individual registries to
cal device, procedure or treatment, and to meet the postmarket surveillance needs for a
describe the natural history of a problem specific manufacturer or a specific product is
or disease. To be useful for device surveil not likely to be efficient or economical. For
lance and assessment of benefits and risks, targeted areas, it may be more cost-effective
registries must contain sufficiently detailed to pursue nationwide medical device reg
patient, device and procedural data, and be istries focused on certain product areas of
linked to meaningful clinical outcomes. high importance as reflected by a large pub
lic health need, patient exposure, uncertain
Use of registries vary. For example, they may long-term or real-world device performance,
be voluntary or designed to meet FDA- or societal cost. For other device areas where
mandated device postmarket surveillance the benefit-risk profiles are well-understood,
requirements. Because well-designed regis registries may not be needed.
tries provide valuable, unique insights into
device performance and device-associated The FDA believes that registry development
clinical benefits and risks, the FDA has en in targeted product areas can both enhance
couraged the development of several device- public health and be cost-effective for in
specific registries and currently participates dustry, health care providers and payers. To
in more than a dozen registry efforts across a foster the development of medical device
number of device areas involving cardiovas registries in key product areas, CDRH will
cular, orthopedic, ophthalmic, and general be convening registry experts and key stake
surgery products. holders, including representatives from
national and international registries, for
For example, in 2011, the FDA helped to Medical Device Registry public workshops6
facilitate the creation of the International to discuss how registries might voluntarily:
Consortium of Orthopedic Registries that
consists of 29 registries from 14 nations and • Leverage experience and expertise to facil
captures data from more than 3 million or itate registry development and initiation;
thopedic procedures.
• Establish common demographic, clinical,
It is neither practical nor feasible to have procedural and device-based data elements;

6 http://www.fda.gov/medicalDevices/newsevents/Workshopsconferences/default.htm


• Develop and share methodological containing longitudinal clinical outcomes
tools for privacy-protected data collec would collectively provide more informa
tion, linking to longitudinal outcomes tion than either system alone and allow for
and analysis; more robust queries and de-identified data
responses. Some health systems, such as
• Enhance interoperability with EHRs Kaiser Permanente and the Veterans Health
and claims data; Administration, have successfully integrat
ed registries into their EHR systems.
• Enhance incorporation of registry data
into EHRs and EHR data into registries; 3. mOderniZe adverSe event rePOrting
and analySiS
• Develop criteria that would render a
registry automatically eligible to sup The FDA monitors postmarket device-re
port an FDA-required post-approval lated adverse events and product problems
study (voluntary certification); through both voluntary and mandatory re
porting to detect signals of potential public
• Create sustainable business models; health concern. Because of the limitations
of spontaneous reporting systems, mod
• Identify priority medical device types ernization of adverse event reporting and
for which the establishment of a longi analysis is a key requirement of a compre
tudinal registry is of significant public hensive medical device postmarket surveil
health importance, such as a subset of lance system. Several ongoing or proposed
Class III or permanently implantable activities will significantly enhance our sur
Class II medical devices; and, veillance capabilities.

• Adopt registry governance structures 3.1 Development of Automated Adverse
that promote rigorous design, conduct, Event Reporting Systems
analysis, reporting of key findings, and
transparency. We are working with partners to explore
automated adverse event reporting systems
Ideally, EHRs and claims data will one day that would facilitate the submission of de
routinely include UDIs. These data sourc vice-related adverse events and minimize
es, therefore, may be used to complement the effort required by the reporter. For ex
and supplement data in device-specific or ample, we are working with 20 hospitals
disease-specific registries. For example, a from our MedSun Network to develop soft
device-specific registry linked to an EHR ware capabilities to export real-time adverse


event data with device identifiers from receives 70 percent of MDRs electronically,
hospital incident reporting systems. The significantly reducing data entry costs of pa
Adverse Spontaneous Triggered Events Re per reports and increasing the expediency
porting (ASTER) study demonstrated that of receiving reports. We anticipate that, ul
facilitated, “triggered reporting” increased timately, electronic reporting of MDRs will
the number of adverse events reported by account for close to 95 percent of all reports.
clinicians. CDRH is piloting ASTER-D to
facilitate the use of hospital EHRs and In 3.3 Develop a Mobile Application for
cident Reporting Systems to detect and au Adverse Event Reporting
tomatically report select device associated
adverse events to the FDA. We recognize the evolving societal role of
mobile applications, their convenience and
The creation of systems that facilitate trig the potential role they may play in improving
gered or automatic reporting of selected public health. CDRH has partnered with
device-related adverse events via the EHR Boston Children’s Hospital in the develop
to the FDA as part of a clinician’s normal ment and implementation of a mobile app
work flow is likely to increase the number for securely reporting medical device adverse
and quality of adverse event reports, de event reports. It is anticipated that this tool
crease under-reporting, and more regularly will facilitate the submission of voluntary re
alert the FDA of potential device-related ports by health care providers and patients.
concerns. The FDA will continue to explore Work is currently underway to refine the app
the development of automated or facilitat so that we may receive such reports through
ed adverse event reporting. FDA’s electronic submission gateway.

3.2 Increase the Number of MDRs 3.4 Modernize the Medical Device
Received Electronically Adverse Event Database

Electronic reporting of device-related ad Reports of adverse events involving medical
verse event reports enhances timeliness, devices received by CDRH are contained in
quality and efficiency of both reporting and the Manufacturer and User Facility Device
postmarket surveillance. CDRH’s eMDR Experience (MAUDE) database (http://
voluntary electronic reporting system pro www.accessdata.fda.gov/scripts/cdrh/
vides the capability for electronic data entry cfdocs/cfmaude/search.cfm).7 The database
and processing of MDRs and utilizes the contains voluntary reports received since
Health Level Seven (HL7) Individual Case June 1993, user facility reports received
Safety Report standard. Currently, CDRH since 1991, distributor reports received since


1993, and manufacturer reports received using automated, computerized statistical
since August 1996. Online capability methods to discover patterns of associations
permits searches for information on medical or unexpected occurrences (i.e., “signals”)
devices that may have malfunctioned or in large databases. CDRH has been explor
caused a death or serious injury. ing the use of these methods in its medical
device adverse event reporting databases to
After 20 years of in-service use, the num systematically prioritize MDRs for CDRH
ber of adverse event records contained in evaluation and review.
MAUDE is now exceeding its design capac
ity. Although CDRH instituted a volun Such methods offer a systematic, automat
tary electronic reporting program in 2007 ed, and practical means of analyzing large
in which we receive adverse events directly datasets and improves efficiency by focusing
in electronic form, the base technology be signal detection efforts on key reporting as
hind MAUDE is antiquated and is unable sociations. Importantly, it offers the poten
to handle the volume and complexity of de tial to identify possible safety issues more
vice reports. In addition, MAUDE cannot quickly than traditional signal detection
take advantage of more modern streams of methods by providing statistically robust,
adverse event reporting (i.e. mobile apps, automated data assessments that can also
EHRs, registries, etc.) and the database account for potentially confounding factors
platform cannot be extended further. and adjust for chance observations.

Therefore, a new adverse event reporting To complement these report and data-driven
system is needed. The FDA is working to efforts, we are developing semantic text
develop a new system, the FDA Adverse mining techniques, which facilitate the au
Event Reporting System (FAERS), with ex tomated extraction and analysis of the narra
panded capacity and modern analytic capa tive text in large numbers of electronic docu
bility for identifying and extracting relevant ments. Using these methods, we are building
information in automated fashion. In addi a computerized search, retrieval and analysis
tion, the new FAERS system will accommo system to quickly and systematically extract
date receipt of adverse event data in more and evaluate information from our adverse
versatile reporting formats. event reporting databases. These efforts will
improve our ability to detect adverse trends
3.5 Rapidly Identify Safety Signals in device performance earlier, minimize
patient exposure to under-performing prod
Safety signals can be more rapidly identified ucts and maintain patients’ privacy.

7 mAuDe does not include reports made according to exemptions, variances, or alternative reporting requirements.


4. develOP and uSe new methOdS fOr assure confidentiality of proprietary in
evidence generatiOn, SyntheSiS and formation, preserve intellectual property
aPPraiSal rights, and maintain transparency are cur
rently being explored and developed.
The evolution of health-related electronic
records, registries and adverse event report However, as discussed previously, medical
ing, as well as the increasingly global nature device postmarket surveillance has unique
of product development and marketing, features that demand customized method
demands the strategic development of in ological approaches. The development of
novative methodological approaches for new tools and methods to generate, syn
evidence generation, synthesis and appraisal. thesize, and interpret postmarket informa
tion will improve the efficiency and qual
In 2010, CDRH launched the Medical De ity of decision-making by identifying new
vice Epidemiology Network (MDEpiNet) and better ways to leverage existing data
Initiative motivated by the need to develop sources by providing more timely informa
and apply innovative methodological strat tion about the benefits and risks of mar
egies to address gaps in studying medical keted products, and by translating data into
devices.8 The development and application knowledge to help better inform regulatory
of novel techniques to collect, analyze, syn and clinical decisions.
thesize, and communicate knowledge about
medical devices can potentially reduce the Selected ongoing or proposed methodologi
burden and cost of postmarket surveillance, cal approaches include:
facilitate the premarket development and
evaluation of new products, and improve 4.1 Quantitative Decision Analysis to
the timeliness, quality and efficiency of Evaluate Benefits and Risks
postmarket decision-making by the FDA,
the medical device industry, health care CDRH is exploring the use of quantitative
professionals, and the American public. decision analysis to help evaluate benefits
and risks of medical devices. This statistical
Some approaches are broadly applicable to research method can be used to better quan
all medical product areas, including medical tify benefits and risks in an explicit and con
devices, and are being developed as part of sistent way both before and after medical
the FDA’s Sentinel Initiative. For example, devices are marketed. Quantitative decision
efforts to protect personal health informa analysis can provide a reliable mechanism
tion, maintain data security and integrity, for incorporating patients’ views on benefit

8 http://www.fda.gov/medicalDevices/scienceandresearch/epidemiologymedicalDevices/
medicalDeviceepidemiologynetworkmDepinet/default.htm


and risk into the assessment of clinical trial parties through the use of clear standards
data, and can characterize and clarify bias for protecting patient privacy as well as
and uncertainty in decision-making. Quan standardized data element names, defini
titative decision analysis can also provide tions and formatting rules. Data standards
context and enhance transparency to health often include information describing proce
care providers, industry and the public re dures, implementation guidelines and usage
garding our regulatory decisions, and pro requirements. Standards facilitate electron
vide information that will help individuals ic reporting, data transfer protocols, data
make important health care decisions. sharing, and data quality.

4.2 Evidence Assessment by Combining The adoption of standard data exchange and
Data from Diverse Data Sources vocabulary specifications and an effective
master data management plan that includes
The ability to combine medical device patient privacy protection would facilitate
performance and privacy-protected clinical the secure receipt, storage, access, and analysis
outcome data from diverse sources would of high-quality data. Notably, data standards
significantly improve the efficiency of post- must meet the needs of the FDA as well as
market surveillance. Such efforts, for example, external stakeholders. The FDA will con
may permit linking of short-term, detailed tinue to work with stakeholders, including
procedural registry data with longitudinal industry, academia and international stan
outcomes data from a claims or administra dards organizations, to promote adoption of
tive database or EHRs. Similarly, several effective data standards.
small data sources may be combined using
meta-analytic and other methods to generate We have begun developing a formal evi
a more accurate assessment of the benefits dence synthesis methodology framework
and risks of a device or class of devices. that is capable of combining de-identified
data from disparate sources including clini
Combining data from disparate sources is cal trials, observational studies, patient reg
made easier and more straightforward by istries, published literature, administrative
the development of common data standards and claims databases, and other external
and a strategy for master data management data sources. This framework will augment
that allows seamless access to consistent traditional regulatory tools and allow us to
high-quality data without the need to per have a comprehensive, up-to-date, benefit-
form duplicate data entry or manipulation. risk profile of a specific medical device at
Data standards promote the efficient shar any point in its life cycle so that optimally
ing or exchange of information between informed decisions can be made and useful


information can be provided to practitio whether a finding represents a “safety sig
ners, patients, and industry. nal” and whether it warrants further investi
gation can be challenging.
4.3 Automated Signal Detection
Factors that may influence the decision in
Increased use of near real-time large data clude the strength of the signal, whether or
bases, such as EHRs or registries, offers not the signal represents a new finding, the
promising opportunities. Modern statisti clinical importance and potential public
cal software and techniques can compare health implications of the issue, and the po
device performance and clinical outcomes tential for preventive measures to mitigate
among marketed products and identify the adverse public health impact.
early signals of concern. For example, the
Data Extraction and Longitudinal Time As part of the CDRH 2012 Strategic Priori
Analysis (DELTA) system has been applied ties, we committed to develop a comprehen
to cardiovascular device registries by the sive framework for the timely evaluation
FDA and academic investigators to estab and management of significant postmarket
lish proof-of-concept for detection of safety signals. Such a framework will consist of
signals of approved cardiovascular devices. several phases including signal detection,
While automated signal detection software risk assessment and signal prioritization,
can facilitate the identification of potential signal refinement, signal verification, and
ly underperforming products, the review signal action. Key components of the signal
and validation of the findings remain criti evaluation and management framework
cal aspects of their use. will address issues of data transparency,
stakeholder participation and timeliness of
4.4 Refinement of Processes for Signal public communication.
Detection and Management

A “safety signal” is information that
arises from one or more sources, and
suggests a new, potentially causal asso
ciation, or a new aspect of a known asso
ciation, between a medical device and an
event or set of related events. The aim
of signal detection is to identify promptly
possible unwanted or unexpected effects
associated with a product. The decision of


bOx c-2.
cOntraSting the current and future StateS Of medical device
POStmarket Surveillance*

Example: Evaluation of Potential Safety Signal
CDRH becomes aware of unexpected adverse events associated with the use of a permanent
neurological implant resulting in the need for early reoperation to have the implant
replaced. After reviewing the available scientific data including published literature, adverse
wevents submitted via its Medical Device Reporting system and information provided by
manufacturers, CDRH cannot determine whether adverse events are occurring at a higher
than expected rate and whether the poor outcomes are associated with one specific product
or with all products of this type.

Current System: CDRH orders all companies who make this type of neurological implant
to conduct a postmarket surveillance study (“522 study”) to determine whether their
product is associated with an increased reoperation rate. Because of the lack of existing
infrastructure and data collection capabilities, it is many months before the first patient is
enrolled in the required study, and typically more than two years before the needed data are
collected and available for analysis.

Future State: Relevant health care systems voluntarily access their EHRs containing unique
device identifiers (UDIs) at the request of the FDA, and, working with the FDA, they help
to determine in weeks rather than years that the increased reoperation rate is associated with
a single model of a single manufacturer’s product. The problematic device is recalled and the
other products remain on the market. Companies whose products are performing well are
not required to conduct additional, expensive postmarket studies.

Example: New Device Receives FDA Approval — Postmarket Study Needed
CDRH approves a premarket approval (PMA) application for a novel implantable stent
in the vasculature to treat arterial narrowings (stenosis) on the basis of a clinical study
following several hundred patients that demonstrated patients receiving the stents are less
likely to require vascular surgery within 12 months after stent implantation. Limited data on
longer term follow-up are available at the time of approval.

Current System: The FDA, as a condition of approval, orders the company to complete a
study of several hundred patients with 5-year follow-up to verify that the real-world, long-
term benefit-risk profile of the device remains the same as that seen in the shorter-term
premarket clinical study. The company organizes a multi-center, post-approval clinical
registry — similar to other multi-center, post-approval registries sponsored by other
companies with similar products. Because other similar products are available and patients

(continued)

* examples are hypothetical and not intended to represent any specific product.
they assume that all regulatory requirements have been met.


bOx c-2. (cOntinued)
cOntraSting the current and future StateS Of medical device
POStmarket Surveillance*

can get the device outside of a research protocol, the company has trouble recruiting
patients to participate in their registry.

Future State: During routine patient care, information about the newly approved stent,
including its unique device identifier, is automatically incorporated into a National Vascular
Stent Registry linked to a claims database, which the company can use instead of making a
registry of their own. The data provides long-term (5-year) follow-up information including
the rate of surgery among patients who receive the stent.

Example: Identification of Potential Safety Signal
CDRH approves several permanent cerebral (brain) implants to treat stroke patients.

Current System: CDRH monitors the MDRs it receives concerning the cerebral implants
but many of the adverse event reports are incomplete or contain insufficient information to
determine if the observed number of adverse events is higher than expected.

Future State: Automated surveillance software provides near real-time analysis of existing
databases (registries or electronic health records) to monitor similar products, identify
potentially underperforming ones and report de-identified data on product performance.
Potential signals identifying poorly performing products are further evaluated to confirm
the initial findings.

Example: Facilitating the Expansion of Labeled Indication
CDRH approves a medical device to treat male incontinence following prostate surgery.
The manufacturer believes the device will also be effective for treating patients with
incontinence due to other conditions.

Current State: The manufacturer is advised to conduct another clinical trial to demonstrate
that the device is effective in treating patients with incontinence due to causes other than
prostate surgery.

Future State: Analysis of de-identified EHR data containing unique device identifiers
demonstrates that within the practice of medicine, physicians have been treating patients
with incontinence due to other causes — and the data demonstrates the device is as effective
as it is in patients with incontinence due to prostate surgery. The company submits the
analysis and CDRH approves an expansion of the labeled indication solely on the basis of
the collected postmarket data.

* examples are hypothetical and not intended to represent any specific product.
they assume that all regulatory requirements have been met.


d. cOncluSiOn

Postmarket surveillance of medical devices presents unique challenges. Although the United
States has a robust postmarket medical device surveillance system, we believe our system can
be strengthened by implementing four key changes to our existing program. It bears empha
sizing that modernizing medical device postmarket surveillance is a long-term effort. Our
proposed strategic changes are intended to complement our existing programs.

CDRH is committed to strengthening our Medical Device Postmarket Surveillance System
to collect, analyze and act on medical device postmarket performance information. We rec
ognize that our postmarket vision cannot be implemented or achieved by the FDA alone.
We have set out this draft plan as a first step and invite e-mail comments on our website and
active participation at our September 2012 public meetings. We welcome e-mail comments
and feedback on this proposal and encourage other ideas and suggestions on how we can
strengthen our existing medical device postmarket surveillance system.


aPPendix
current Selected medical device POStmarket authOritieS

CDRH’s postmarket authorities include Study (PAS) for a device in a premarket
the following : approval application (PMA) order, or
by regulation at the time of approval
• Medical Device Reporting — CDRH has or subsequent to approval of the de
the authority to require mandatory medi vice. Post-approval requirements may
cal device reporting (MDR) from device include as a condition of approval of
manufacturers, user facilities and import the device, continuing evaluation and
ers.9 Manufacturers, user facilities and periodic reporting of device safety, ef
importers must report under the MDR fectiveness and reliability.11 The FDA
regulations whenever they become aware states in the PMA approval order the
of an event that reasonably suggests that a reason or purpose for such requirement,
device may have caused or contributed to a the number of patients to be evaluated
death or serious injury. In addition, certain and the reports required to be submit
malfunctions must be reported. Failure to ted. There are no time limitations (e.g.,
comply with the MDR requirements will length of study) for a PAS. Failure to
render the device “misbranded,”10 and may comply with a PAS constitutes grounds
result in the issuance of an Untitled Letter, for withdrawal of approval of a PMA.12
Warning Letter or more severe penalties
such as injunction, seizure or civil money • Postmarket Surveillance — CDRH
penalties. In addition, health professionals has the authority to order a Postmarket
and consumers may voluntarily report to Surveillance Study (often referred to as
the FDA adverse events relating to the use “522 studies”) for certain Class II and
of marketed medical devices. III devices, generally for a duration of
up to 36 months.13 Postmarket surveil
• Post-Approval Studies — CDRH has lance may be ordered if:
the authority to order a Post-Approval

9 section 519 of the fD&c Act; 21 cfr part 803.
10 section 502(t)(2) of the fD&c Act.
11 21 cfr 814.82.
12 21 cfr 814.82(c).
13 section 522 of the fD&c Act. the study duration may exceed 36 months if the manufacturer agrees to extend the study
timeframe, or if no agreement can be reached, after the completion of a dispute resolution process. if fDA’s order is for
a device that is expected to have significant use in pediatric populations, the study duration may exceed 36 months if
such period is necessary in order to assess the impact of the device on growth and development, or the effects of growth,
development, activity level, or other factors on the safety or efficacy of the device.


■ device failure would be reasonably • Recalls — CDRH has the authority
likely to have serious adverse health to issue a "cease distribution and noti
consequences; fication" order, if, after providing the
■ the device is expected to have signifi appropriate person with an opportu
cant use in pediatric populations; nity to consult with the agency, CDRH
■ the device is intended to be im determines that “there is a reasonable
planted in the body for more than probability that the device would cause
one year; or, serious, adverse health consequences or
■ the device is intended to be a life- death.” The order may require the ap
sustaining or life-supporting device propriate person to immediately:
used outside a device user facility.14
■ cease distribution of the device;
The FDA may also order postmarket ■ notify health professionals and de
surveillance as condition of clearance of vice user facilities of the order; and,
a device that is expected to have signifi ■ instruct these professionals and
cant use in pediatric patients.15 device user facilities to cease use of
the device.
• Registration and Listing — CDRH has
the authority to require establishments CDRH may then amend the order to
that are involved in the manufacture, require a recall of such a device under
preparation, propagation, compound certain conditions.17 A mandatory recall
ing or processing of medical devices order from FDA will include, among
intended for commercial distribution others things, provisions for notifica
in the United States to initially register tion to individuals subject to the risk
and thereafter register annually, and to associated with the use of the device. If
submit device listing information. The a significant number of such individuals
required listing information includes, cannot be identified, FDA may notify
among other things, a list of all com such individuals under section 705(b)
mercially distributed devices being of the FD&C Act.18
manufactured or processed at the estab
lishment and, in certain situations, the • Device Tracking — CDRH has the
labeling for such devices.16 authority to require a manufacturer to

14 section 522 of the fD&c Act.
15 section 522(a)(1)(b) of the fD&c Act.
17 section 518(e) of the fD&c Act; 21 cfr part 810.
18 21 cfr 810.13(d). under section 705(b) of the fD&c Act, fDA may disseminate information regarding devices in
situations involving “imminent danger to health, or gross deception of the consumer.”


adopt a method of tracking for a Class II believe that the device was not
or Class III device, if the device meets one properly designed or manufactured
of the following three criteria and FDA with reference to the state of the art
issues an order to the manufacturer: as it existed at the time of design or
manufacture;
■ the failure of the device would be ■ there are reasonable grounds to
reasonably likely to have serious believe that the unreasonable risk
adverse health consequences; was not caused by failure of a person
■ the device is intended to be im other than a manufacturer, importer,
planted in the human body for more distributor, or retailer of the device
than 1 year; or, to exercise due care in the installa
■ the device is a life-sustaining or tion, maintenance, repair, or use of
life-supporting device used outside a the device; and,
device user facility.19 ■ notifying device users would not by
itself be sufficient to eliminate the
CDRH’s tracking authority is intended unreasonable risk, and the repair,
to ensure that the tracked device can be replacement, or refund is necessary
traced from the manufacturing facility to eliminate such risk.21
to the person for whom the device is
indicated (i.e., the patient).20 • Withdrawal of PMA Approval and Re
scission of 510(k) Clearance — CDRH
• Repair, Replace, Refund — CDRH has has the authority to withdraw a PMA
the authority to order the manufacturer, approval22 and to rescind a 510(k) clear
importer, or distributor of a device to re ance, under certain circumstances.23
pair or replace it or to refund its purchase
price if, after affording an opportunity • Ban Devices — CDRH has the author
for a hearing, CDRH determines that: ity to ban a device, by regulation, which
would prevent it from being legally
■ the device poses an unreasonable risk of marketed, if the agency determines, on
substantial harm to the public health; the basis of all available data and in
■ there are reasonable grounds to formation, that the device presents a

19 section 519(e) of the fD&c Act; 21 cfr part 821.
20 21 cfr 821.1(b).
21 section 518(b) of the fD&c Act.
22 515(e) of the fD&c Act; 21 cfr 814.46.
23 the fD&c Act does not expressly address rescission of a device clearance. however, agencies have inherent authority
to reconsider their decisions in certain circumstances, such as where there has been fraud or error, or to correct their
mistakes. See, e.g., American Therapeutics, Inc. v. Sullivan, 755 f. supp. 1, 2, (D.D.c. 1990).


substantial deception or unreasonable “adulterated” and/or “misbranded,”
and substantial risk of illness or injury and/or will constitute a “prohibited
which cannot or has not (after request) act” under the FD&C Act.28 CDRH
been corrected or eliminated by labeling has authority to initiate enforcement
or by a change in labeling, or by a change actions, including seizure of devices,
in advertising if the device is a restricted injunctions, civil money penalties, and
device.24 The procedures for banning a criminal penalties for violations involv
device are described in 21 CFR Part 895. ing adulterated or misbranded devices
and for prohibited acts.29
• Restrict Devices — CDRH has the au
thority to, by regulation, restrict the sale, • Removal of a Device as a Predicate —
distribution, or use of a device if, because Under Section 513(i)(2) of the FD&C
of its potentiality for harmful effect or Act, a device may not be found to be sub
the collateral measures necessary to its stantially equivalent to a predicate device
use, there cannot otherwise be reasonable that has been removed from the market
assurance of its safety and effectiveness.25 at the initiative of the Secretary or that
A regulation may restrict the device only has been determined to be misbranded
to be sold upon receiving the oral or or adulterated by a judicial order.
written authorization of a practitioner
licensed by law to administer or use such
device, or upon other conditions as pre
scribed by the regulation. If prescribed in
the regulation, the label must include ap
propriate statements of any restrictions
required by such regulation.26 CDRH
also has the authority to restrict PMA
devices by requiring as a condition of
approval that the device be restricted.27

• Enforcement Actions — Failure to com
ply with certain postmarket require
ments for devices will render the device

25 section 520(e) of the fD&c Act.
26 section 520(e) of the fD&c Act.
27 section 515(d)(1)(b)(ii) of the fD&c Act; 21 cfr 814.82(a)(1).
28 sections 301, 501, and 502 of the fD&c Act.
29 chapter iii of the fD&c Act, prohibited Acts and penalties.


food and Drug Administration
center for Devices and radiological health
10903 new hampshire Avenue
silver spring, mD 20993
http://www.fda.gov/medicalDevices
http://www.fda.gov/radiation-emittingproducts

Fda postmarket surveillance-report-september-remediated

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

Andere mochten auch

Andere mochten auch (6)

Ähnlich wie Fda postmarket surveillance-report-september-remediated

Ähnlich wie Fda postmarket surveillance-report-september-remediated (20)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Fda postmarket surveillance-report-september-remediated